1 This work was first published in 1843 and went through eight editions, the last appearing in 1872. All my references are to John Stuart Mill, A System of Logic Ratiocinative and Inductive, 8th edition, edited by J.M. Robson with an introduction by McRae, R.F. Collected Works of John Stuart Mill, vol. 7 (Toronto: University of Toronto Press 1974)Google Scholar. Hereafter this work is designated by The initials SL.

2 Laudan, Laurens ‘Theories of Scientific Method from Plato to Mach,’ History of Science 7 (1969) 31Google Scholar

3 Losee, John A Historical Introduction to the Philosophy of Science (Oxford: Oxford University Press 1977), 148Google Scholar: ‘Mill's philosophy of science is an example of the inductivist point of view,’ which contains ‘certain extreme claims about the role of inductive arguments both in the discovery of scientific laws and in the subsequent justification of these laws.’ Ryan, Alan John Stuart Mill (New York: Pantheon Books 1970),Google Scholar xi: Mill is the ‘author of a philosophical system … [of] inductivism.’ Nagel, Ernest ed., John Stuart Mill's Philosophy of Scientific Method (New York: Hafner Press 1950),Google Scholar xxxviii: Mill's Logic, summarily, is ‘an account of induction.’

4 For example: Anschultz, R.P. The Philosophy of J.S. Mill (Oxford: Clarendon Press 1969);Google Scholar Britton, Karl John Stuart Mill (Harmondsworth: Penguin Books 1953);Google Scholar and McCloskey, H.J. John Stuart Mill: A Critical Study (London: Macmillan 1971).CrossRefGoogle Scholar

5 This essay was published in 1859. References (indicated hereafter by the initials OL) are to the following edition: John Stuart Mill, On Liberty in Essays On Politics and Society, Robson, J .M. ed., Collected Works of John Stuart Mill, vol. 18 (Toronto: University of Toronto Press 1974).Google Scholar

6 Ryan, Alan J.S. Mill (London: Routledge & Kegan Paul 1974), 138Google Scholar

7 Feyerabend, Paul ‘Democracy, Elitism, and Scientific Method,’ Inquiry 23 (1980) 7CrossRefGoogle Scholar

8 Ibid., 4; cf. Feyerabend, Paul ‘How To Defend Society Against Science,’ reprinted in Hacking, Ian ed., Scientific Revolutions (Oxford: Oxford University Press 1981) 158–60.Google Scholar

9 In his System of Logic and elsewhere Mill exemplifies the widespread disregard by Victorian philosophers of Hume's thought. Interest in Hume arose late last century only after the ‘edition of his works by T.H. Green and T.H. Grose (1874)’: Passmore, John A Hundred Years of Philosophy, 2nd ed. (Harmondsworth: Penguin Books 1966),Google Scholar 531 note 1. Of Hume's philosophy, Passmore says, ‘Mill knew practically nothing’: Ibid., 13. For differences between Mill and Hume's attitudes to induction see R.F. McRae's introduction to SL, xxxv.

10 Mill's quotation is from Whewell, William History of Scientific Ideas, vol. 1 (London: Parker and son 1858), 65.Google Scholar

11 At one place Mill speaks uncharacteristically of ‘absolutely certain’ inductions, but in the third (1851) and following editions this is toned down to ‘quite certain and quite universal’ inductions (SL, 322 text and notes).

12 The idea of fallibility is not confined to this particular passage which Mill added to the final edition of the Logic. In all editions of the work Mill is explicit about the fallibility of the inductive procedures of science (SL, 568), and about the uncertainty of scientific predictions. Because of this uncertainty he requires that laws of nature be expressed in the form of tendencies, not invariant relations (SL, 443-45, 570, 575 note t).

13 In this capacity the methods systematically develop the principle of Francis Bacon that discovery of uniformities among the confused assemblages of circumstances in nature requires variation of the circumstances of experiments.

14 See also, John Stuart Mill, Autobiography in Autobiography and Literary Essays, Robson, John M. and Stillinger, Jack eds., Collected Works of John Stuart Mill, vol. 1 (Toronto: University of Toronto Press 1981), 217.Google Scholar

15 Mill's category of ‘unknown’ causes, the subjects of speculative hypotheses, are defined by him in terms of the knowledged possessed by scientists at any given time. Such causes may be either observable or unobservable. Although Mill seems to be at least dimly aware of the importance in science of unobservable putative causes (vortices, gravitation, etc.), he fails to explain by what indirect means (direct empirical inspection being out of the question) scientists are to comply with his requirements and, so, establish (a) these causes’ existence and (b) whether they are ever present as the antecedent conditions of experiments.

16 Mill, Autobiography, 167Google Scholar

17 Ibid., 167, 189-91, 215-17

18 Mill, John Stuart ‘The French Revolution,’ London and Westminster Review 5 and 25 (July, 1837) 48Google Scholar

19 There are, to be sure, other possible sources of Mill's hypothetical method, including Herschel's Preliminary Discourse on the Study of Natural Philosophy, which he read (and reviewed) in 1831. But Mill himself points out that it was only after reading Whewell in 1837 that he returned to Herschel's work and properly digested its ideas (Mill, Autobiography, 217Google Scholar). Further, we must note that while Mill ascribes hypotheses important rules in economics and geometry in his writings before 1837, they are not hypotheses in the sense of conjectures or suppositions, but rather idealizations or approximate truths resultant from inductive generalizations (SL, 224, 901-3).

20 The Earlier Letters of John Stuart Mill 1812-1848, Mineka, Francis ed., Collected Works of John Stuart Mill, vol. 12 (Toronto: University of Toronto Press 1963), 363Google Scholar

21 Mill, Autobiography, 217 note

22 Ibid., 255 note

23 Mill quotes from Auguste Comte, Cours de philosophie positive, 6 vols. (Paris: Bachelier 1830-42), 2: 434.

24 Mill never wavered in holding that enumerative induction is the method of discovering and validating the law of universal causation, simple pre-scientific generalizations, and basic laws of mathematics. For this and related points see: SL, 287 note, 304, 506, 562, 568, 609, and see also: Mill, John Stuart Auguste Comte and Positivism (Ann Arbour: University of Michigan Press 1961),Google Scholar 46 note.

25 For affirmations of this view in two other works see: Mill, Comte and Positivism, 55-6, and John Stuart Mill, ‘Grote's Aristotle,’ in Essays on Philosophy and the Classics, Robson, J.M. ed., Collected Works of John Stuart Mill, vol. II (Toronto: University of Toronto Press 1978), 485.Google Scholar

26 It is unclear to what ‘fashion’ Mill refers in this passage. In part at least it may be to Comte and his followers. Comte is the only ‘contemporary thinker … [to be] criticized by name’ in Liberty: Himmelfarb, Gertrude On Liberty and Liberalism (New York: Alfred A. Knopf 1974), 90.Google Scholar

27 This is not to deny that there occur in the Logic some scattered, uncharacteristic remarks which presage On Liberty's method of criticism. In the fourth edition (1856) and in all later editions, to take one example, Mill opines that ‘many errors’ or false hypotheses ‘have been got rid of by coming to a knowledge of facts which were inconsistent with them’ (SL, 433). This and the other presagements, along with the critical method of On Liberty itself, seem to bespeak the influence of Whewell. Mill pertinently records that Whewell's ‘account of the manner in which a conception is selected, suitable to express the facts, appears to me perfectly just. The experience of all thinkers will, I believe, testify that the process is tentative; that it consists of a succession of guesses; many being rejected, until one at last occurs fit to be chosen’ (SL, 297 my emphasis: cf. 304, 432, 503).

28 My interpretation of On Liberty in no way entails that Mill revoked the philosophy of social science which he developed in detail in book 6 of the Logic. In Liberty, I am holding, his epistemic concern is with theories and doctrines that are non-empirical in character and which therefore lie outside the ambit of science. Mill never ceased believing that there are factual laws of social life, whose discovery by social scientists would help to improve the efficacy of government policies for social reform.